Title
Giuliana Giusti (Venice): Determiners and (un)interpretable nominal features

CASTL Colloquium

Date:  21.09.2006

Time:  19:15

Venue:  E010 1

Description
1. A number of imperfect parallels between noun phrases and clauses are to be observed:

a.Both noun phrases and clauses project an argument structure BUT in nominals argument structure appears “less” obligatory (Grimshaw 1990)
b.Both have a “subject” which must respect the hierarchy of the theta grid BUT in nominals it appears to be optional (Cinque 1980).
c.Both can be claimed to have three layers (a lexical layer, an inflectional layer, and a complementation layer) all of which are split according to a universal hierarchy BUT nominals display less rich a structure.
d.In both we find structural case BUT in sentences we typically find two (nominative and accusative) in nominals we typically find one (genitive) if any.
e.In both we find discourse driven displacements BUT the left periphery of noun phrases is much more difficult to detect (Giusti 1996, 2005, 2006).
From these I conclude that nominals and clauses share the same properties imposed by the general system (including EPP feature) but differ on some property to be reduced to the nominal vs clausal nature of the extended projection.

2. I reduce this difference to a single formal feature [Reference]. Clauses typically have Temporal reference. Noun phrases typically have (object) Reference or Denotation (e.g. Person, Number, and in some cases Gender). Being related to temporal reference, finiteness is only found in clauses. Being related to denotation, Number is only found in noun phrases. From this, I conclude that it is not the lexical-semantic features but the formal features distinguish the two linguistic objects.

3. The following step extends Rizzi’s split CP hypothesis to the DP. The conjecture is that while the splitCP contains Force and Fin the split DP contains Case and Number.
a.Force expresses the discourse value of a matrix clause. But in an embedded clause it realizes selectional requirements of an external lexical head. Matrix noun phrases are rare (but not impossible, cf. noun phrases used as exclamations, or a vocatives). The typical occurrence of a noun phrase is however embedded into a clause. I propose that the counterpart of Force is Case, which realizes the requirements of an external head.
b.Fin is part of the highest layer of the clause (the CP, in which interpretable features are valued) but at the same time it is directly linked to a lower head T in which the inflectional morphology of the verb is merged/checked. In a parallel fashion Number is part of the highest layer of the noun phrase (the DP in which the interpretable features are valued) but it agrees with lower heads in which Num is shared with adjectival modifiers.

4. I adopt Pesetsky and Torrego’s (2004) proposal of a bipartite distinction: [_/val], [i/uF] yielding four possible combinations: [_iF] (unvalued interpretable), [_uF] (unvalued uninterpretable), [valiF] (valued interpretable], [valuF] (valued uninterpretable) and claim that:
a.Parallel to T, which is interpretable in clauses and uninterpretable in noun phrases (cf. Pesetsky and Torrego 2004), Num (bundled with person) is interpretable in noun phrases and uninterpretable in clauses.
b.Parallel to instances of [_iT] (unvalued interpretable T) in the intermediate layer of the clause, we find intermediate instances of [_iNum] in the intermediate layer of the noun phrase.
c.These instances of unvalued formal features can be bundled with other formal features (of any of the four types above).

5. In the rest of the paper I go back to the imperfect parallels above and derive them from the properties of [Tense] vs. [N], [valuF] (valued uninterpretable) and claim that:
a.Parallel to T, which is interpretable in clauses and uninterpretable in noun phrases (cf. Pesetsky and Torrego 2004), Num (bundled with person) is interpretable in noun phrases and uninterpretable in clauses.
b.Parallel to instances of [_iT] (unvalued interpretable T) in the intermediate layer of the clause, we find intermediate instances of [_iNum] in the intermediate layer of the noun phrase.
c.These instances of unvalued formal features can be bundled with other formal features (of any of the four types above).

5. In the rest of the paper I go back to the imperfect parallels above and derive them from the properties of [Tense] vs. [Num].